Hiring people is like making friends. Pick good ones, and they'll enrich your life. Make bad choices, and they'll bring you down. Who you work with is even more important than who you hang out with, because you spend a lot more time with your workmates than your friends.

Wait, why am I talking about hiring? Isn't unemployment stubbornly high? Aren't tons of folks unable to find new work? That's certainly the case for many industries, but not ours. In fact, our company's job board -- which lists positions for programmers and designers across our industry -- has more help-wanted postings than ever. We recently hired two new people. Something is going on.

I'd like to share a bit about how we go about hiring at 37signals. Hiring is something we rarely do -- we're intentionally small at 20 people -- but we've developed a method that has worked very well for us. It allows us to find the right people and keep them happy. In 11 years, only two people have left the company -- and one recently returned after working elsewhere for seven years. (Welcome back, Scott!)

So, what do we do? First, we hire late. We hire after it hurts. We hire to alleviate pain, not for pleasure. Who hires for pleasure? Any company that hires people before it needs them is hiring for pleasure. It's an indulgence we've never allowed ourselves.

We're happy to skip over the perfect catch if we don't have the perfect job for the person to do. Right now I know there are some great designers and programmers available, and I'd love to have them on our team. But we don't have any openings.

I've run into a lot of companies that invent positions for great people just so they don't get away. But hiring people when you don't have real work for them is insulting to them and hurtful to you. Great people want to work on things that matter. Inevitably, a great person working on imaginary work will turn into an unsatisfied person. Then he'll leave.

It's easy to say, "There's gotta be stuff you'd like to do if you had more people." And, of course, there is stuff I'd like to do. But I believe it's good to operate at the limits of your organization. Limits force you to come up with creative, elegant solutions. Being forced to get more done with fewer resources is the right kind of pressure.

A smaller team keeps you focused. It crowds out all the things you'd like to do and replaces them with the things you have to do. It forces you to prioritize and focus on the next most important thing instead of the next "wouldn't it be cool if…" thing. There are always plenty of those.

How do you know if you really need someone? A good rule of thumb is this: Have you already tried to do the job yourself? If you haven't done the job, you don't really understand the job. Without that fundamental understanding, it's hard to judge what constitutes a job well done.

For example, a few years back, we decided it would be a good idea to bring on a business-development person, someone who could follow up on partnership inquiries and other new business opportunities we were being pitched. Up to that point, we'd pretty much been ignoring those e-mails. We were just too busy doing other things.

So we began interviewing people. Some were very qualified and had great references. But because we hadn't actually followed up on these new business opportunities before, it was hard to know exactly how to proceed with a candidate. In the course of conducting job interviews, we quickly learned that because none of us had even tried to pursue unsolicited partnership requests, none of us could evaluate a candidate's skills appropriately. Would the candidate be good at doing something we know nothing about? How would we even know?

So after meeting with a variety of people, we stopped the search and began looking into these inquiries ourselves. It quickly became obvious that most of these deals weren't worth pursuing anyway. If we hadn't taken that extra step, we might have hired someone to spend time on something we didn't even want. That's definitely not good for us -- and it's not good for a biz-dev person's career, either.

We've learned this lesson with other positions, too. Before we hired our first customer service person (Sarah), I did all the customer service, about two years of answering e-mails. David, my business partner, and Jamis, one of our programmers, did all of our system administration before we hired our first system administrator (Mark). We found great people because we thoroughly understood the jobs.

Once we begin vetting candidates, we also behave a little differently. For one thing, we ignore resumés. In my experience, they're full of exaggerations, half-truths, embellishments -- and even outright lies. They're made of action verbs that don't really mean anything. Even when people aren't intentionally trying to trick you, they often stretch the truth. And what does "five years' experience" mean, anyway? Resumés reduce people to bullet points, and most people look pretty good as bullet points.

What we do look at are cover letters. Cover letters say it all. They immediately tell you if someone wants this job or just any job. And cover letters make something else very clear: They tell you who can and who can't write. Spell checkers can spell, but they can't write. Wordsmiths rise to the top quickly. Another rule of thumb: When in doubt, always hire the better writer.

We look for effort, too. How badly does this person want the job? Pestering is not the same as effort, though. We hired a designer named Jason Zimdars because: 1. He was good, and 2. He made more effort to get the job than anyone else. He built a special website pitching his skills just for us. So few people make the extra effort like Jason did. (Check it out to see what I mean: jasonzimdars.com/svn.)

During interviews, we love when potential hires ask questions. But all questions aren't equal. A red flag goes up when someone asks how. "How do I do that?" "How can I find out this or that?" You want people who ask why, not how. Why is good -- it's a sign of deep interest in a subject. It signals a healthy dose of curiosity. How is a sign that someone isn't used to figuring things out for him- or herself. How is a sign that this person is going to be a drain on others. Avoid hows.

We also try to test-drive people before hiring them full time. We give designers a one-week design project to see how they approach the problem. We pay them $1,500 for their work. Sometimes, we'll hire someone on a contract basis for a month to see how we feel about the person and how the person feels about us. Sometimes that project is just a few hours a week because the candidate already has a day job. But that's often enough to check out the person's work, how the person communicates, and how the person works under pressure. These real-work tests have saved us a few mismatched hires and confirmed a bunch of great people.

Finally, we never let geography get in the way. We hire the best we can no matter where they are. We're based in Chicago, but we have programmers in Idaho and California, system administrators in North Carolina and downstate Illinois, designers in Oklahoma and Colorado, a writer in New York City, and others in Europe. This obviously wouldn't work for customer-facing folks, but for most everyone else, it does. The best are everywhere. It's up to you to find them.

Jason Fried is co-founder of 37signals, a Chicago-based software firm, and co-author of the book Rework.